Purifi Class D

[Miska]

12 hours ago, barrows said:

Yeah, I suspect there is some level of efficiency where more headroom will not bring improvements, but without trying one never knows for sure.  My own speakers are 4 ohm, and rated 90 dB for 1 watt, but I suspect that rating is a bit generous, as the speakers do appear to like some more power.  Theoretically, the 400 Watts at 4 ohms should be much more than enough, but the amp with even more headroom has sounded better in the past.

 

Usually it is the loudspeaker's complex impedance that matters from amplifier perspective. Not the absolute resistive impedance power. And that combined with amplifier's damping factor.

 

If you don't have complex impedance curves of the loudspeaker, minimum EPDR gives a good idea how difficult the loudspeaker is to drive. Something like 2.5 ohm EPDR is not unusual, and these appear usually around mid bass frequencies for conventional dynamic speakers. And around treble area for electrostatic speakers (electrostatic is essentially a capacitor, higher the frequency, closer it gets to short-circuit). Magnetostatic speakers have their own unique behavior.

 

High damping factor vs minimum EPDR is best comparison if you want to get idea of how much the amplifier has control over the speaker.

 

If you have electrostatic, it is essentially a short-circuit for class-D's switching frequency.

 

[Miska]

On 11/23/2022 at 4:37 PM, 57gold said:

Who is producing high powered/high current Purifi amps?

 

I believe latest NAD M-series?

 

[Miska]

On 7/15/2023 at 6:06 PM, Jud said:

If read carefully, the article doesn’t actually say Fourier analysis is insufficient

 

It is good to remember though that human hearing can beat Fourier time-frequency analysis though. All the information persists through the transform though, but just goes through "unnoticed".

 

[Miska]

3 hours ago, mocenigo said:

Sorry, but what you just wrote does not make sense at all. Literally. If *all* information passes through the transform, how can human hearing beat it? And isn’t it the purpose of the transform of just changing the domain of the representation of the data and not alter it? What does it “notice” this information mean in this context?

 

Well, @Jud already answered to that. But when you increase frequency resolution by making the transform longer, you lose on the time resolution because the transform covers longer section of the time. And vice versa. Since human hearing is not based on Fourier transform, it is not limited in this way either, it can detect frequency and timing independently.

 

The information is there, in the data, but it gets statistically distributed such way that it doesn't appear, but gets "lost in the noise".

Just like in statistics, there is lot of information there, but whether some information appears in the statistical analysis is another matter.

 

[Miska]

3 hours ago, mocenigo said:

Also, some people seem to think that the Fourier transform is fundamental to how sound is represented. This is wrong, it is perfectly possible to have a chain from audio recording to reproduction that does not perform any FT or FFT, except for the purpose of analysing the content. There are formats based on the FFT but they are optional. And, on top of it, the lossless ones also carry the difference between the signal and the deconvoluted version; so this is not an issue anyway.

 

But you can add the transform to the data path and it wouldn't alter the data at all. You wouldn't be able to tell whether it is there or not.

 

[Miska]

7 hours ago, mocenigo said:

True. But the effect of ringing is overstated, in fact it is essentially absent from "true musical signal," or inaudible, as long as the source does not clip and is properly bandwidth limited. In fact, I never heard of a single study proving it is audible if these conditions are met.

 

It is not absent, it is already created in the ADC as result of that band-limiting. I have demonstrated this also in practice.

 

7 hours ago, mocenigo said:

The sigma-delta modulation is not to enable a "simpler, less expensive final analog reconstruction filter", it is a conversion process and in fact it requires a more significant investment analog reconstruction filter.

 

R2R PCM requires much more complex and expensive analog filter to properly remove the images. And the amount of oversampling you can do is limited by settling time issues, because conversion output needs to settle within 1/2 LSB within fraction of the sample period. Otherwise lot of error (distortion) is generated.

 

7 hours ago, mocenigo said:

So, digital filtering without upsampling will not remove the conversion artifacts, i.e. the images. They are a product of the conversion and they depend on the sampling rate.

 

Yes, there always needs to be analog filter, but digital filter is needed to help the analog filter do it's job.

 

7 hours ago, mocenigo said:

No processing of the signal will reduce them.

 

Yes it does, they get further apart and lower in level due to transfer function of sample-and-hold process. This makes it easier for the analog post-filter to remove what ever remains.

 

We can use Holo Audio NOS DAC combined with oversampling digital filter front-end as an example.

 

Without digital filter, running at 44.1 kHz, 0 - 22.05 kHz sweep output looks like this, you get wide spread of images, reconstruction accuracy about 4 bits:

 

Same source, but with oversampling to 705.6 kHz it looks like this, still one image around 705.6k visible, reconstruction accuracy about 13 bits:

 

And with oversampling to 1.4112 MHz it looks like this, first image would be around 1.4112M at about -110 dB, reconstruction accuracy about 18 bits:

 

 

When we oversample to 11.2 MHz, the first image is at almost 10x higher frequency... If we oversample to 45.1584 MHz (DSD1024) it is way way much further away.

 

 

In addition, combining oversampling with suitable word length and noise-shaper designed to linearise conversion stage, we can correct low level linearity errors inherent to R2R...

 

Here's output at oversampled to 705.6k, 24-bit TPDF dithered output:

 

Then when with engage suitable noise-shaper and set word length to 20-bit, we can see the distortion components are gone:

 

[Miska]

47 minutes ago, mocenigo said:

I wrote more investment, i.e. there is some R&D that is more significant.

 

Yes, I've been doing my current project since 1998. So 25 years of R&D investment has been going to it.

 

But it is still more economical per playback system.